Methods: An instrument was created based on a literature review of appearance criteria used to assess autologous breast reconstructions. Criteria included in the rating scale were positioning, defects, projection, shape, inframammary fold, medial contour, and overall appearance. Postoperative photographic images of 32 patients who had previously undergone TRAM flap breast reconstruction were evaluated by 4 medical judges (2 microsurgery fellows and 2 Breast Center nurses at Barnes-Jewish Hospital) and by the patients themselves. In addition, the patients rated their overall satisfaction. The evaluations were performed twice, one month apart, by the judges and patients.

Results: Test-retest reliability for each aesthetic item in the scale was better for patients than for judges (weighted kappas 0.53 - 0.87 vs. 0.46 -0.63). The instrument demonstrated slightly better overall internal consistency when used by patients (Chronbach's alpha value of 0.92) than by judges (alpha 0.85 - 0.91). In addition, for patient evaluations, all individual aesthetic criterion scores correlated well with the overall appearance self-score (Spearman's rho 0.55 - 0.89). For judge responses, most, but not all, individual aesthetic criteria correlated with the overall appearance scores (rho 0.50 - 0.73); positioning, inframammary fold, and medial contour scores correlated less with the overall judge scores (rho 0.28 - 0.38). Patient ratings of overall appearance and of overall satisfaction correlated highly with each other (rho of 0.81). Overall inter-rater agreement on each criterion was poor among judges, as well as between judges and patients (weighted kappa -0.07 - 0.35). Finally, no correlation was found between the judges' perceptions of any individual criteria and patient overall appearance or satisfaction scores (rho -0.19 - 0.56).

Conclusions: The instrument demonstrates a slightly better internal consistency and reproducibility when used by patients than by medical professionals. Expert evaluations did not correlate well with patient evaluations, and the poor inter-observer agreement among the judges creates a potential problem for interpreting independent judge scores. The data suggest that patient input regarding item-specific criteria should definitely be included in evaluations of breast reconstructions. Ratings of the same set of patient images are currently being performed by a panel of reconstructive surgeon judges and will be included in the presentation.